Environment understanding by multiple feature-based region segmentation

This paper presents a recognition method for finding meaning of artificial and natural objects in the image. It defines natural and artificial objects for recognizing region in the image. It separates and extracts various objects in the image about the natural and artificial. It acquires campus images using CCD camera that mounted on robot and extract meaningful features. It classifies natural and artificial objects using various region segmentation methods. In the campus environment, the objects have characteristics of color, shape and geometric information. The recognition method segments the defined objects using color, shape and geometric information of each object. To acquire feature object, image information selected according to characteristics object. The recognition method presents an understanding of outdoor environment by extracting representative objects among natural and artificial objects. It confirmed the usability of presented analysis method in outdoor environment using restricted environment like a campus.